This study aims to produce computer-based learning media with a scientific approach to probability material. The development model used is the 4D model which consists of the stages of define, design, develop, and disseminate. The instrument used was a validity instrument in the form of a validation sheet and a practicality instrument in the form of a student response questionnaire. The data collection techniques used were validation sheets and student response questionnaires. The data analysis technique used in this research is quantitative data analysis and qualitative data analysis. Validation is carried out by four validators who are experts in their fields. A limited trial was conducted on 16 students of SMP Mutiara Harapan Pangkalan Kerinci. Students are divided into two groups, consisting of students with high, medium, and low abilities. Each group tries two learning media. From the results of data processing, it was found that the average validation of learning media on the media aspect was 3.47 with very valid criteria and the material aspect was 3.53 with very valid criteria. This means that all aspects of the device in the learning media developed are related consistently. While the average percentage obtained from the student response questionnaire in the limited trial was 90.59% with very practical criteria. This means that the learning media developed can be used in learning activities. It can be concluded that the learning media developed are valid and practical.